State Institutions, Civic Associations, and Identity Demands

Front matter

AMY H. LIU JOEL SAWAT SELWAY

TATE INSTITUTIONS i CIVIC ASSOCIATIONS

EGIONAL MOVEMENTS in Greater Southeast Asia

Part 1. State Institutions I

Singapore’s Entry and Exit, 1963-65

Elvin Ong

State Institutions in North Taiwan versus South Taiwan: Hokkien Language Recognition

Chun-Ying Wu and Amy H. Liu

State Institutions in North Sumatra, Indonesia: National Identification of the Bataks Risa F. Toba

Contrasting the Onset of National Rebellion Mary Anne S. Mendoza-Davé

State Institutions in East Malaysia versus West Malaysia: Containing Grievances in an Ethnoreligious Dominant Party System

Mohamed Salihin Subhan and Kai Ostwald

State Institutions in Northeast Thailand: Lao Ethnics and the Thai Identity Jacob I. Ricks go vi + Contents

Part 2. Civic Associations

7 Civic Associations in East Timor, Indonesia: Lessons from ‘Timor-Leste’s Independence Bids S. P. Harish

8 Civic Associations in Bali, Indonesia: Coastal Reclamation and Ethnic Mobilization Ryan Tans g Civic Associations in Northern Myanmar: Pan-Ethnic Nationalism and Sub-Ethnic Mobilization in Kachin State Alexandre Pelletier

10 Civic Associations in Maluku, Indonesia: Explaining the Failure of the South Maluku Republic Movement Jessica Soedirgo

11 Civic Associations in North Sulawesi, Indonesia: Ethnic Politics and the Level of Mobilization Problem for Bangsa Minahasa Jeremy Menchik

12 Civic Associations in North Thailand: Cultural and Regional Movements in the Lanna Region Joel Sawat Selway

Part 3

13 Conclusion: Necessary Conditions, Context Conditionality, and Democratic Theory Amy H. Liu and Joel Sawat Selway

Commentary. Regionalism in Comparative Perspective: A Cardinal Breakthrough in the Study of Ethnicity, Secession, and Political Violence Henry E. Hale

References Contributors

Digital materials related to this title can be found on the Fulcrum platform via the following citable URL: https://doi.org/10.3998/mpub.12333333

IQI

2.1 3.1 3.2 4.1

Figures

Declining First-Generation Waishengren Population National Pride by Ethnic Groups

Batak Representation in Indonesia’s Cabinet

The Process: From Colonial Education Policy

to Rebellion Onset

Penetrative Education Policies

Shallow Education Policies

KJA-Controlled and Contested Areas in Kachin State (2011-19)

10.1 Claimed Territories of the South Maluku Republic

48 58

69 76 77 77

172 181

Tables

Ethnic Demographics by Region (1957—60)

Ethnic Groups in Taiwan

Ethnicity of Respondent’s Father (1993)

Ethnicity of Each Parent (Respondents Born Pre-1950) Ethnicity of Each Parent (Respondents Born Post-1950) Language Used to Conduct Interview (Hokkien Speakers) National Identity of Respondents

National Identity of Respondents (Hokkien Sample) Isan Language ‘Treatment Effects

Select Demonstrations

ForBALI’s 13 Reasons to Reject Reclamation

Official Declaration of GEMPAR Benoa Bay

Ethnic Groups in North Myanmar (Kachin State) Minahasans in Comparative Perspective

Governors of North/Central Sulawesi, 1960-2000

The 12 Core Values of the Thai People

Overview of State Exclusion Meets Civic Inclusion

(in This Volume)

Acknowledgments

In a project of this ambition—covering almost twenty regions in eight countries—the thanks spread far and wide. We first want to thank the incountry experts and informants who made each of the individual chapters possible. As a community of scholars, we are constantly reminded that we are students of the citizens and cultures of the countries that we study. We express deep gratitude to them for allowing us access to their worlds and for imparting wisdom to us in its variety of forms.

Next, we want to thank each of the contributors to the volume and the additional scholars whose chapters did not quite fit into the final organization of the book. Mai Truong and Paul Schuler have an excellent account on whether there is a north-south identity divide in Vietnam. The account can be found in a special issue of Asian Politics and Policy (31, no. 1 [2021]). We would have loved to include that article in this volume, and we encourage all to read it. Additionally, Shelli Israelsen wrote an interesting account on the lack of mobilization among Bamar and Karen women, which also unfortunately did not make it into the volume.

Earlier drafts of all these chapters were presented at a 2019 American Political Science Association Mini-Conference (“Explaining Identity Formation in Asia”). We would like to thank the six discussants for their comments—both on the individual chapters and on the overall framing for the book: Johanna Birnir, Kikue Hamayotsu, Allen Hicken, Sarah Shair- Rosenfield, Dan Slater, and Meredith Weiss. We also feel privileged that Henry Hale made the time to offer a commentary. It is our hope that our efforts to explain the different types of regional movements can be applied beyond Southeast Asia.

xii + Acknowledgments

We must, of course, impart thanks to Elizabeth Demers and Haley Winkle at the University of Michigan Press and to Dan Slater, editor for the Emerging Democracies series. They shared our passion for the project and offered guidance throughout the process. The project also benefited tremendously from two anonymous reviewers—not something we always get to say with honesty and genuine appreciation.

Finally, we would like to thank the Southeast Asia Research Group (SEAREG) for its vision and the community it has built. All of the contributors are SEAREG members. An edited volume such as this one—which takes rich insights from area studies knowledge and applies them to important theoretical debates in the field—is exactly the type of scholarship the organization emphasizes. Even with COVID interrupting the pace of the project, the ties established through SEAREG led to continued fruitful exchanges on the ideas of the book over the past four years.

Amy H. Liu and Joel Sawat Selway

Southeast Asia has been home to some of the world’s bloodiest separatist conflicts. Myanmar alone has the most simultaneous ongoing civil wars; it also has the /ongest ongoing conflict between the Karen National Liberation Army (KNLA) and the Myanmar government. Beyond Myanmar, in other parts of the region, violent ethnoregional movements continue to rage, from Western Papua (Irian Jaya) in Indonesia to Mindanao in the Philippines to the Malay-Muslim provinces of Thailands Deep South. Meanwhile, numerous other ethnic insurgencies across Southeast Asia have dissipated, been crushed, or been resolved, such as the Hmong guerilla resistance in Laos and the Timorese insurgency in Indonesia.

At the same time, peaceful separatist movements also abound across Southeast Asia. In the Cordillera region of the Philippines, the government made provisions on two separate occasions for referendums on regional autonomy, though the proposal failed both times. In the Lanna region of Thailand, peaceful movements for separatism appeared briefly in 2013-14, characterized by small protests and banners being hung from bridges. In Malaysia, organizations like the Sarawak Association for Peoples’ Aspiration and the Borneo Heritage Foundation advocate for independence. And in Vietnam, the Khmer Krom and Montagnards have proposed independent states.

Some of these separatist movements have links to regional political parties. However, regional parties that champion regional interests—without advocating for separatism—are even more numerous. Examples include the majority of Sabah and Sarawak parties in East Malaysia. Despite their history with armed insurgent groups, many ethnic parties in Myanmar today are not demanding independence. The Mon Unity Party, for example, advocates for the representation of Mon cultural interests. This was the case both before and after the coup. Most Acehnese parties in Indonesia (Aceh being the only region not held to the national party registration requirements) likewise do not advocate for separatism. Relatedly, there are sharp divides between the North and the South in Taiwan and Thailand— with parties essentially representing one particular region even if they do not have the region in their name.

Lastly, most regions across Southeast Asia have experienced no regional movements at all. In Thailand, the Pak Tai region (South), Khorat region, and Northern Khmer area (Buriram, Sisaket, and Surin) have made no agitations; even the Isan region has had nothing as open as the separatist activities in Lanna in 2013-14. In Malaysia, Johor, Kedah, Negeri Sembilan, and Terengganu—all of which speak different vernaculars within the Malay language family—have witnessed no movements, just whispers of rumors from time to time. The same goes for the majority of the regions in the vast and diverse archipelagos of Indonesia and the Philippines. The north, central, and southern regional divides of Laos and Vietnam— despite significant historical differences, distinct cultural heritage, and clear factions within the Communist parties of their respective countries— have experienced no regional movements. Cambodia has also notably been absent from any discussion of regionalized political movements thus far.

In short, we see a remarkable amount of diversity in regional political movements across Southeast Asia, ranging from (1) nothing at all in the majority of regions to (2) the formation of regional parties targeting group-based interests, from (3) parties explicitly advocating for autonomy or independence to (4) movements inciting violent separatist rebellions. What explains why identity-based political demands emerge in some regions but not others? How do these demands transform into political movements? And what explains the form these regional movements take?

Regions, Regionalism, and Regional Movements

In this book, we use the term region as the unit of analysis rather than “ethnic group” for several reasons. First, “ethnic group” is a broader term used in the discipline to identify a group with a “myth of collective ancestry” (Horowitz 1985). This can include groups based on race, language, religion, and region. Thus, regional groups are just one type of ethnic group. While regional groups often are differentiated by other characteristics as well, not all ethnic groups are defined by a geographic region. Many ethnic groups are not geographically concentrated—for example, the Chinese in Thailand, Malaysia, and Indonesia (Liu and Ricks 2022). Note that our research question is on the mobilization of nationalist causes, which by definition is political control (sovereignty) by a geographically defined territorial group—that is, a region (Gellner 1983). Thus, it makes no theoretical sense for all ethnic groups to be in our universe of cases. Instead, we should only examine the geographically concentrated ones. To consider ethnic groups regardless of their spatial spread artificially inflates the number of null cases.

Second, it is sometimes the case that regional groups are not considered ethnic until they have engaged in identity-based violence. In the United States, for example, most regions are not considered ethnic groups. Yet, if we were to consider which regions are possibly ethnic, we suspect many would say “the South” or “Texas”—based precisely on their history of secessionism and a strong cultural identity. This definitional trap is precisely what we want to avoid. Not all geographically concentrated ethnic groups turn to separatism. To consider ethnic groups only when they have mobilized ignores the null cases to the detriment of theory.

Third, it is also sometimes the case that regional movements are diverse but share one common denominator. The chapter on North Myanmar shows that the regional movement—composed of multiple linguistic groups—has mobilized under a larger Kachin identity. Yet ironically, because the identity has mobilized via the Kachin Baptist Church network, it has been so broad that the ethnic identity borders on being meaningless. Put differently, it is also possible that what would otherwise look like a monolithic ethnic movement is scuppered by the fact that the region itself is internally diverse on another dimension—as was the case of South Maluku in Indonesia. To consider only ethnic groups, then, fails to theoretically acknowledge the internal diversity of regions and the political process by which they form broader ethnic identities within the region, a process of which the conclusion is not foregone.

Finally, it is also possible that ethnic identities are endogenous to the region. It is only after the process of mobilizing for regional autonomy that a distinct ethnic identity emerges. An example of this is in the South Philippines. While there was a large Muslim population even during the Spanish period, the ethnoreligious identity did not manifest until after independence. ‘To consider only ethnic groups risks empirically truncating the mobilization process. In short, we believe region is theoretically and empirically the most appropriate unit of analysis.

We use the term regionalism to refer to the politicization of one’s regional identity. We treat region as distinct from linguistic identity, both of which can be treated as a subset of the broader way in which social scientists define ethnicity—that is, an “identity categor[y] in which membership is determined by attributes associated with, or believed to be associated with, descent” (Chandra 2006, 397). The key difference between region and language dimensions of ethnicity is, of course, the necessity of association with a place of origin. In other words, there is a geographic area that determines membership in the ethnic group. This definition, however, begs conceptual clarification due to the ex post nature in which scholars routinely treat politicized regional activity. Virtually every study on secessionism—whether it is violent separatism or peaceful independence movements—examines regions in which these phenomena have already manifested. Moreover, this literature almost exclusively treats the ethnic group as the unit of analysis. However, the very identification of a region as an ethnic group is often determined by virtue of the fact that they have made calls for independence.

Consider Thailand, for example. Thailand is generally divided into five cultural regions. There is Central Thailand and four periphery regions: North, Northeast, South, and Extreme South. While there is a huge literature on the civil war in the Extreme South, it is devoid of comparison to the other regions because of its “ethnic distinctiveness.” Indeed, the Extreme South is often labeled the Malay-Muslim region. Similarly, in the post- 2001 election era—after intense protest activities unfolded throughout the Northeast region—people began labeling the Northeast region as the Isan or Thai-Lao region with a common explanation for the social unrest emphasizing linguistic distance between the Lao and Thai speakers. Then in early 2014, as protesters hung separatist banners from bridges in the North region, scholars suddenly “discovered” the Lanna ethnic group and its long history of separateness from historic “Thai” states. Thus far, the South has not been depicted so much in ethnic terms. This is not because of a lack of ethnic distinctiveness. In fact, the linguistic distance between Southern Thai and Central Thai is just as large as that between the Isan (Northeast) and Lanna (North) groups. Instead, the absence of any ethnic overtones is because the region has been largely on the same political side as the Central Thais.

In short, there is a tendency among scholars to consider a region as an ethnic group only when we witness a politicized identity. We need an empirical strategy that is free of this and the other selection biases outlined above. Specifically, we need to be able to identify ex ante which areas of land constitute a region. On the one hand, if we go too micro—for example, a city block—we are left with an infinite number of regions that give us no analytical traction. On the other hand, if we go too macro—for example, the river delta—we risk having regions that have no substantive meaning. In short, we must find some middle ground between these two extremes. But we must do so while avoiding selecting “obvious” regions where there is perceived cultural uniqueness and/or politicized regional identity.

Let us walk through an exercise as an example. In Malaysia, the East- West divide is a commonly cited geographical classification. East Malaysia is the northern part of the island of Borneo, separated by hundreds of miles from West Malaysia—a narrow strip of peninsular land extending from Thailand in the north to Singapore in the south. Government statistics are routinely calculated comparing these two areas. In East Malaysia, we have seen some political movements advocating independence, not just separately for Sabah and Sarawak—that is, the two provinces there based on historic ethnolinguistic groups—but for the entire land mass. However, to date, we have not seen such appeals based on the entirety of the more populous West Malaysia.

This raises the question of whether there is a West Malaysian regional identity. On the one hand, if we dismiss West Malaysia as a region, this brings us back to East Malaysia: Why is Sabah and Sarawak lumped together as one region? On the other hand, if we do think of West Malaysia as a region, it also ignores how the peninsula itself was divided by states based on historic, politically independent sultanates. And each one of these states constitutes a regional identity in much the same way that Sabah and Sarawak do.

And as it turns out, this discussion is not specific to the East-West divide. We see the same issue with other aggregations between single states and the entire peninsula. For example, we can talk about a Northern Peninsula Malaysia collectively: Kedah, Kelantan, and Terengganu. But does the north also include Penang? While Penang is in the north, it is different in ethnic and religious makeup, voting patterns, and history of incorporation by the British. So, if we exclude Penang and leave it as its own region, this forces us to justify why Kelantan, Terengganu, and Kedah—three separate states—are lumped together as one. In short, what we see here is that the number of potential arrangements based on purely geographic space is infinite.

This is problematic. If we are interested in why some regions develop movements but others do not, we need to be clear on the universe of cases. Imagine if we had drawn four concentric circles from the geographic center of Malaysia and called each area a region. Not only would these regions be theoretically irrelevant, but empirically they would also be an empty set: There is no evidence or belief of association. One way to scale back the list of possible regions is to emphasize a key part of the definition: an attribute associated with, or believed to be associated with, descent.

Here, we use the term “descent” to distinguish this from selection based on ethnicity in its traditional sense. Many regions have strong identities that they would feel uncomfortable labeling an ethnic identity. Take the region of Western Thailand. This region was centered on the city-state of Phetchaburi, which maintained high levels of autonomy within Siam. Linguistically, they have a local variant of Central Thai. However, religiously and administratively they were closely aligned with the Siamese center of power. The very suggestion to people in this region that they are of a different ethnicity than those in Bangkok (the capital) would seem preposterous. Nevertheless, at some base level they identify with the idea of a Western Thai region; they recognize descent from a particular set of provinces and kinship groups within those provinces. It is precisely regions like this that get lost in an ethnically focused selection strategy. And it is possible that the likes of Western Thailand represent the majority of regions in the world, but theoretically we ignore them.

Association with descent based strictly on geography is rare. This is in part because there are often other identity attributes associated with the area—whether it is language, history (i.e., political autonomy), color, race, tribe, nationality, religion, socioeconomic status, or even partisan support. Consider the regional identities in Northeast Malaysia or South Taiwan. Both areas have long been associated politically with the political opposition. While the key is to establish association with descent based on region of origin, such claims may be hard to separate from other identity attributes. Note, however, that it is not necessary for this project that we isolate the different markers as long as the region remains the focus of the theory.

Let us return to our regions of Malaysia. While it would be difficult to classify West Malaysia—in aggregate—as a regional category, it is not wholly impossible. We could argue there is a peninsular identity based on class (it is more metropolitan and has higher levels of wealth); likewise, a case could be made on language, race, and/or religion. What we would need is to be able to demonstrate that the peninsular identity is meaningful in some way. Here, the best evidence, though extremely weak, is to turn to linguists’ categorization of peninsular Malay languages/dialects as distinct from the Bumiputera languages on East Malaysia. Having met our minimal definition of region, however, it then becomes an empirical question why a West Malaysian identity has failed to be politically mobilized.

We can make a better case for Northeast Malaysia (minus Penang). The three states in this region (Kedah, Kelantan, and Terengganu) have a shared history as a part of the Siamese Kingdom. And as a result, their vernaculars—while each different—have Thai linguistic influences. Likewise, their brand of Islam is more conservative. And finally, they share similar voting behavior: This region is a stronghold for the Malaysian Islamic Party (PAS).

Alternatively, a strong case can be made for looking at the individual states as regions. After all, each state has its own unique Malay language, independent political histories, and associated cultures. This strategy is certainly easy and consistent; after all, every country (with the exception of Singapore) has subnational states/provinces. The challenge, however, is that these subnational states could very well be the endogenous product of some political processes—for example, tax collecting purposes.

In sum, while we may be able to identify conceivable regions at different levels of aggregation, the region must possess a minimal level of descent-based identity. Given the preceding discussion, in this volume we conduct our selection process using the most exogenous approach possible. First, we take existing lists of ethnic groups in each country and eliminate those that are not geographically concentrated. This gives us most of our groups in Indonesia, whose fairly unique archipelagic structure tends to be an exception to many of the selection biases discussed above. Next, we append to this list regional groups. We consider geographic areas in each of the nine cardinal regions in each country: north, northeast, east, southeast, south, southwest, west, northwest, and center. This is just our starting population of possible regional groups. We seek to identify some minimally verifiable basis of a regional identity, which is often related to the existence of a historic political center in the region. Where we find evidence of a regional identity, we keep it; where we do not, we remove it—leaving us with what is our working list. Note that strength of identity is not an attribute of inclusion. Moreover, we allow regional identities to overlap; for example, Western Thais are part of the larger Central Thai region as well. We discuss our empirical strategy in greater detail below.

Given this strategy, we end up covering a variety of regions. Our sample includes an ethnoreligious region in North Sulawesi, Indonesia (Minahasan); a historical ethnonational region in South Malaysia (Singapore); an ethnoracial region in North Myanmar (Kachin); and an ethnolinguistically homogeneous region in South Taiwan. With a definition of regionalism not determined by saliency, we can select different regions in which political movements do not emerge—enabling us to build theories on regional saliency and identity politicization.

What Do We Know about Regional Movements?

There has been a lot of work done on regions that have developed movements at the extreme: violent separatist groups. The same, however, cannot be said for movements somewhere in the middle—and certainly not the case for movements that have had no identity-based demands whatsoever. In the remainder of this section, we review the literature both specific to Southeast Asia and to comparative politics more generally.

Regional Movements in Southeast Asia

Most qualitative analyses of separatist movements focus on the cases where demands are already widespread and where significant action has been taken. A/ the literature on separatism in Southeast Asia falls under this label. The majority of the analyses tend to focus on one single group within the country—for example, the Acehnese in Indonesia or the Moros in the Philippines. And as such, this body of scholarship is dominated by country-level units of analysis (e.g., Brown and Ganguly 2003). What this means is that explanations tend to also be at the national level.

Consider one such explanation, government ideology. Liow (2016) argues that the violence we see in Southern Thailand is because of the government’s refusal to recognize the Malay language. The refusal is because accommodation would otherwise conflict with how the government envisions the nation: one that necessarily includes the periphery. In contrast, Malaysia expanded its conceptualization of the dominant ethnic group, Bumiputera, to include those in East Malaysia, other indigenous populations, and non-Malay groups such as the Kedah Thais. This accommodation has translated into limited conflicts. Another country-level explanation focuses on the weakness of the state. Tarling argues that when states are politically stable, “citizens daily live a connection with the material rewards of their imagined community” (2004, 14). There is no reason for separatism. Weak development, however, creates fertile conditions for separatist demands.

Not all explanations are at the country level. For example, Bertrand and Laliberté (2010) look at international-level dynamics. They argue that colonialism created arbitrary boundaries that have made some subnational movements inevitable. Likewise, the Cold War, regional rivalries among the major powers, and the War on Terror have also contributed to why violent separatist conflicts emerge. While these aforementioned volumes are insightful, the focus at the country or international level does not give us the requisite leverage to explain why some regions but not others experience movements—let alone why only some regional movements evolve into violent rebellions.

Reid’s Imperial Alchemy: Nationalism and Political Identity in Southeast Asia (2010) is the only volume we are aware of that undertakes analysis at the subnational level comparing groups within the same country. He examines why the Acehnese—but not the Bataks—have engaged in separatist violence in Indonesia. Both groups are geographically concentrated. Yet, the Bataks have been much more mobile in the post-independence era, being transformed into “predominately urban and diasporic communities, for whom regional or ethnic separateness is no longer imaginable.” Conversely, the Acehnese have remained isolated in the northwest edge of the archipelago—where the “memory of the state” remains strong. To this end, Reid’s approach comes closest to ours. Not only does he focus on groups rather than countries, but he also includes groups with varying levels of ethnic salience.

Nevertheless, we depart from Reid in several respects. First, Reid’s focus on ethnic groups assumes a different universe of cases that potentially includes groups that are geographically dispersed (e.g., Chinese in Indonesia) and groups confined to capital cities (e.g., Indians in Bangkok). Neither of these groups is likely to engage in any of the outcomes we consider in this volume—and certainly not separatism. Second, Reid’s universe of cases risks excluding groups that may be difficult to distinguish from the majority, such as the Pak Tai in Thailand. This region has been part of Siam for hundreds of years, though highly autonomous prior to the 2oth century. In part due to this shared history, but also due to being on

IO * STATE INSTITUTIONS, CIVIC ASSOCIATIONS, AND IDENTITY DEMANDS the same political side as the dominant Central Thai over the past three decades, it would be easy to exclude such a group from an ethnicity-based analysis. However, the Pak Tai would be identified in our selection strategy due to inhabiting the cardinal region of Southern Thailand and having an easily identifiable descent-based identity. Finally, our objective here is general theory building. We are interested in identifying necessary and/or sufficient regional movements, which was not Reid’s goal.

General Comparative Politics Literature

In comparative politics, the focus is almost always on the ethnolinguistic group; itis rarely about the region. Consider the literature on secessionism, which looks at group demands for independence: Why are there demands (Jenne 2004; Saideman 1997); what are the demands (Jenne, Saideman, and Lowe 2007; Sorens 2012); and why are some mobilization efforts successful but others are not (Englebert and Hummel 2005; Horowitz 1981; Islam 1985)? Likewise, the literature on civil wars finds that when groups fight over territory, the conflicts are more violent and longer in duration because land is perceived to be neither divisible nor substitutable (Vasquez 1993; Hensel and Mitchell 2005; Mitchell and Thies 2011; Gubler and Selway 2012). The attachment to the land (Sack 1986; Touval 1972) means that even after the conflict is over, social intolerance is much more pronounced between the groups engaged in territorial conflicts (Tir and Singh 2015). And, finally, consider the literature on ethnic accommodation in democracies. Whether political power should be moderated through majoritarian institutions (Horowitz 1985; see Reilly 1997) or shared through consociational designs (Lijphart 1999; see Lijphart and Crepaz 1991; McGarry and O’Leary 2007), scholars champion federalism as an institution. The intuition is that by giving ethnolinguistic groups political control over cultural matters, this can facilitate democratic consolidation.

Consistent across all three bodies of literature is the focus on the group—with region only in the background as a reinforcing cleavage, as part of the underlying mechanism, or related to a solution to the conflict. This unit of analysis is partly data driven—that is, until recently there has been no datasets on regions (see Hooghe et al. 2016). That choice is itself driven by a theoretical bias that ethnolinguistic groups are where the action begins. This approach, however, ignores the constructivist literature that suggests regions may develop ethnic identities to further their cause— that is, geographical difference could precede ethnic differences.

Additionally, theories on secessionism, civil wars, and ethnic accommodation need to develop mechanisms that emphasize and are preferably unique to geographic space. Consider the recent creation of the Telangana state in India. Formerly part of Andhra Pradesh, Telangana shares the Telugu language with the rest of Andhra Pradesh. Indeed, language was the basis for state creation at independence. However, the Telangana movement emphasized a separate history from the rest of Andhra Pradesh as part of the Hyderabad Princely State. The movement also emphasized inequalities in education and water rights. The idea of a distinct Telangana people is strongly tied to regional mechanisms rather than linguistic or cultural differences.

By focusing on groups instead of region, we risk one of two types of selection effect bias. The first is selection on the dependent variable. ‘The tendency is to look at groups where we see demands—whether it is more cultural recognition or outright political separation. And from here, scholars try to identify some common denominator among the cases where the demands were met. This strategy, however, ignores the instances where demands were absent; it also overlooks the sample of cases where demands were made but not recognized by the national government. Failure to consider these two other scenarios means we can only identify a necessary condition at best. If the non-cases also share the explanatory variables, however, we may miss the real explanation of what is going on. Empirically, we lose valuable information on how often these demands are made and met.